Abstracts

Rationale: Our primary goal is to understand the mechanisms leading to temporal lobe epileptic seizures and to develop more effective treatments. We are studying the influence of synaptic connectivity on neural network activity, with emphasis on the spatial patterns of propagation related to epileptic seizures. Our computer model of the CA3 region of hippocampus allows us to observe network activity patterns and to analyze their relationship to neural and network parameters. We tested the hypothesis that under conditions supporting periodic synchronous discharge of area CA3, reductions in excitatory synaptic transmission could result in repeated patterns of network activity. These repetitive patterns could underlie the local rhythmic EEG activity that precedes focal seizures. This activity would be analogous to the reentrant activity observed in the heart as a consequence of loss of function mutations of cardiac sodium channels.Methods: We tested this hypothesis with our computer model of a neural network (Swiercz et al. J Neuroph 2007;24(2):165-74). This model includes 10,000 pyramidal and 225 interneurons. Since we are modeling CA3 structure, the majority of connections are excitatory recurrent collateral synapses that demonstrate use-dependent depression. We implemented synaptic loss of function by reducing the rate of glutamate release and increasing the time of glutamate replenishment.Results: As predicted from experimental data, use-dependent depression terminated the synchronous population activity in the control network. However, the networks with reduced rates of glutamate release exhibited prolonged network activation. These results are consistent with the in vitro results obtained by Jones et al. (J Neuroph 2007;97(5):3812-8) in which the rate of glutamate release was reduced by replacement of extracellular Ca2+ with Sr2+. The spatial patterns of activity differed dramatically between normal networks and those with reduced rates of glutamate release. In control networks, we observed circular patterns of excitation while spiral-like patterns of activity were observed in networks with reduced glutamate release.Conclusions: The reentrant activity we observed computationally can be attributed to reductions in the degree of activity-dependent synaptic depression. Under these conditions, when circular waves of excitation return to their origin, the origin has already recovered from a activity-dependent depression, so that repeated activation of the same pathway can take place, and spiral-like activity may occur. The following conditions increase the probability of repeated pathway activation: 1) the conduction time along the circular pathway closely matches the time required to recover from synaptic depression 2) glutamate release rates must be sufficiently high to support synaptic transmission, but low enough to reduce the degree of synaptic depression. Importantly, even under these conditions spiral-like waves often terminate upon collision with other spontaneous waves of activity, providing a possible explanation for the relative rarity of seizures vs. interictal spikes.